Posted by: zyxo | August 22, 2008

Nerve cells establish and break down contact


Diagram of neuron with arrows but no labels. M...Image via Wikipedia Contact between nerve cells is also constantly being set up and dismantled in adults.
This means our brain is more active and intelligent than we thought.

I wonder if this thechnique could been used in artificial neural networks (ANN’s) ?
This would mean ANN’s where not only the weights between the connections are adapted, but where connections are established and broken during the learning process.

Reblog this post [with Zemanta]

Responses

  1. Well, certainly it could and it has indeed been used. The only problem with it is that to know when to set up a connection or when to break one requires much more amount of computation if the connection is to connect between two units that are arranged to be quite distant from each other. As for units that are near each other, why not just assume they’re always connected and thus avoid that trouble of breaking and connecting?

    Though it’s true that with enough computation power, in order to make an ANN better what you’ve said is what an ANN must have.

  2. Hyq,
    I am not an expert in ANN’s. I do not understand how to interpret ‘distance’ between neurons. There is not such a thing as physical distance. Perhaps you mean number of layers in between ?
    In the case when you use an algorithm that sets up and break connections that could be done without considering any distance. Unless you give the neurons virtually a physical location as to copy the structure of a biological brain ?


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Categories

%d bloggers like this: